illuminarti

I think that slicers and/or firmware could be more intelligent for sure, but I'm not sure that that thing isn't just more or less the same as acceleration, though, contrary to what the blurb on the page says? Perhaps it isn't described as well as it might be. What is needed, I think, is to decouple the extruder from the x-y motion, so that pressure build up/release is managed. E.g., Finish extruding early on a layer, so the pressure dies down before the z-change or move. Simplify3D has a concept of this called 'Coasting' which works really well. For instance, in the double-wall cup, shown below, I set the coast distance is just over 2mm, leaving an apparent gap in the wall around the z-scar area. The head passes over that gap, but not actively extruding. The pent up pressure oozes into that gap anyway, so it closes it just fine, and then when the head pauses to change layer, there is less scarring as a result. It's a crude control, but quite effective. Marlin does have some code in it that is supposed to try to model and manage the hysteresis effects and pressure in the Bowden tube, and compensate for it, but I'm not sure if it actually works - it's not used in the current UM firmware.

It's odd that there seems to be such a sharp transition at the bottom-left to top right diagonal. Does changing your acceleration speed (Maintenance -> Advanced -> Motion Settings) make a difference to how the pattern comes out?

Welcome!

Here's how we level all the beds on the printers we produce in Memphis: The first thing to do is to make sure that all of your springs are reasonably tight. If they aren't, it's impossible to find the right height with a single sheet of paper, because the tension on the springs is minimal until they're compressed a bit. So, look through the bed from front to back, and adjust the back thumbscrew until the terminal block in the back left corner is about 1mm from touching the lower plate. Tighten the front screws about the same amount, to keep the bed roughly level as a starting point. Then heat the nozzle and make sure it is clean at the tip. Now run the leveling wizard again. When adjusting the rear height, just use the dial on the front of the printer. When adjusting the front corners, use the thumbscrews. I recommend not using the 1mm-then-a-paper-thickness approach. Instead, on both passes level the bed to the point where the nozzle just touches the glass. This is easy to see if you look along the surface of the glass; you can see the nozzle touch its own reflection. When setting each point, move the bed up until it just touches the nozzle tip, then back it off and allow it to settle untouched, and then gently close the gap again. If you find that you cannot compress the front springs enough to get the bed down to where it needs to be, then simply raise the back of the bed a few turns of the thumbscrew, and restart the leveling wizard. You want to end up with all the springs in a middle position, with a gap of about 10-13mm between the two plates of the bed assembly. The springs should be neither totally compressed, nor so loose that they aren't applying any meaningful upward force on the bed. By doing two passes at the same height you should get fewer surprises; the second pass around should only require very minor adjustments. And aiming for the point where the nozzle touches the glass is a much easier target than trying to interpret the feel of nozzle on paper. This will leave the nozzle about 0.1mm too close to the bed (the assumed paper thickness). This is very rarely a problem, if you print with a 0.2 or 0.3 first layer, and helps to ensure good adhesion.

Actually, no... it should give you 0.3 + (7 x 0.1), I agree, but it doesn't. The number of solid layers on the bottom doesn't take into account the first layer thickness. So you get 0.3 + (9 x 0.1) instead. Well, actually, no, you get 0.3 + (10 x 0.1) because there's another bug that you always seem to get one more top/bottom layer than you should. I filed a bug report against CuraEngine for both these things a few days ago.

Bear in mind that these extruders all suffer from back-pressure related under-extrusion to some extent. Given that the Ultimaker² design is fundamentally the same as the Ultimaker Original design in the most important aspects of the physical drive mechanism, I imagine that the Ultimaker² has similar issues - maybe even more so - than the Ultimaker Original did. See my findings with the Ultimaker Original here: http://www.extrudable.me/2013/04/18/exploring-extrusion-variability-and-limits/ Given this, I think it would be interesting to see if extruder tension affects the perceived infill under extrusion, and also whether temperature affects the problem. Also, I wonder whether, for instance, the knurling on the machines that have this problem is noticeably different/worse than that on machines that don't have the issue.

Skanect is a good, fairly cheap program for building 3D scans from Kinect or Primesense/Asus scanners. However, while the quality is pretty good at the resolution of a person, or bigger, it's not really suited for scanning small detailed objects. I can't really imagine the belt scanning very well, tbh. Especially with a Kinect, which is lower resolution that other more dedicated scanners like the Carmine v1.09.

By all means retest the gcode, but I checked Nicolinux's somewhere earlier in this saga, and it was spot on. When it's printing, the printer doesn't know or care what the layer height is - it just does what the gcode says, based on the e coordinates. So thicker layers will have larger e amounts for the same line segments, but the printer has no understanding of how far below the nozzle the previous layer is. Ultigcode uses volumetric e coordinates - cubic mm to extrude - while normal gcode has linear mm of plastic to feed in for the e coordinates. One test that might be interesting is to slice the same piece with ultigcode and regular gcode, keeping all the settings and temp the same, and see if there is a difference between the results - which, if both gcodes look correct, would imply an issue in how Marlin is converting ultigcode volumetric e coords into linear ones.

Z-Brush is an incredible piece of software which I love using, and which I think is actually really good value for money for everything it can do. But it's important to understand that it is solely a freeform sculpting package. You basically start with a ball of clay and pull it into shape. It's not for making accurate flat-sided geometric shapes with known dimensions - like mechanical parts. Sculptris is a very, very cut down version of Z-Brush, and has one major drawback for 3D printing use - it doesn't remesh the shapes - i.e., re-covering your edited object with a new simpler mesh of triangles when it intersects itself or gets twisted - but instead keeps contorting the original one, so that it is fairly easy to end up with a very complicated, self-intersecting mesh that will confuse slicers and be hard to print. DSM is the free version of SpaceClaim - which is another awesome piece of software for doing more geometric/mechanical work. SpaceClaim is a few thousand dollars, but still cheap compared to something like SolidWorks I think. DSM does most of what you need for most 3D print modeling, however, and is free.

It's copper grease that is used as a lubricant at the junction of the steel nut and aluminum plate, to prevent corrosion and ensure that the nut can be turned easily. (And btw, technically, that's not usually called the extruder. It's the printhead or hot end. The extruder is really the black plastic filament drive part on the back of the printer - although to avoid confusion I tend to refer to that as the 'feeder assembly' these days.)

It's very easy to drill the nozzle out to 0.8mm yourself; drill bits are cheaply available on Amazon, and the nozzles are relatively soft metal.

That's really weird. The tip is very flattened and the opening itself looks to have widened. Some sort of chemical reaction of the material and the brass? How does the inlet of the block look? Are there any signs of the inside of the nozzle getting eroded?

Yes, I think this is a known behavior if the printer is connected via USB to a computer - you can get these freeze ups when the computer wakes or sleeps. The short answer is simply 'don't do it'.

I haven't timed it... but I will. But it doesn't sound entirely unlikely to me. Most of the time is going to be heating the bed. You get diminishing returns on heating the bed, the hotter you go. The head probably only takes a couple of minutes, as it's tiny. The bed is much harder. I generally only print PLA, so I don't heat the bed hotter than about 60. That probably takes 4 or 5 mins I'd guess. So adding that extra 30 degrees to the bed, and heating the head... it feels about normal to me. But I'll do some tests.

Yes.

The steps per mm setting is a function of the motor, pulleys and belts. While there can be other factors at play for the extruder, you really shouldn't need to adjust the steps for the x and y axes. You can get some shrinkage as the plastic cools, and there can be some other things going on that leave the final part fractionally too small, depending on its geometry. But 2mm sounds like a lot. Can you provide some examples of what you are seeing?

Sadly, no. We don't have any spare parts for sale. Those have to be ordered from the Netherlands.

No, 0 will make the first layer be the same as all your other layers - 0.06mm. Which is really too thin; it's hard to level the bed that accurately. The settings you posted above show that the first layer height is set to 0.3mm - which is a more reasonable value (Although anything in the 0.15-0.3 range is probably ok). Make sure that the first layer goes down slightly squashed into the bed. If you level too high to start with (then your print will be taller by that amount). Also make sure you are running a recent version of the firmware, as firmwares before 14.07 had some problems with how bed leveling was recorded. Some other things to remember: the finished object is always going to print using whole layers. So the finished height will always be equal to the first layer height, plus (n x layer height) where n is an integer. Secondly, the shell thickness should really be an exact multiple of the nozzle size (0.4mm) or else Cura will do funky things.

I'm sorry that we haven't done a better job of communicating the status to you. I sent you a PM here on the forum with more information.

Reversing the polarity to the fans shouldn't hurt anything, but they will only run with the correct polarity. However, if you shorted out the pins between the board and the fans - got a spark from them most likely - there's a chance that the something broke on the electronics board. (I know that if the LEDs get shorted, the dimmer circuitry can blow and get stuck with the fans 'half on' - I presume the same can happen for the fans). So first check all the wires again, and make sure all the connections are good. Turn the fans on full, and see if they run. Try kickstarting them by hand to see if they are getting tool little power). Try and measure the voltage at the board pins (should be 24V), and also check the wiring at the printhead.

Nicolinux - Power the printer off, and put the head in the middle of the bed. Then move it left to right by hand, back and forth, and look carefully at the left and right sliding blocks that aren't moving (much). I imagine that you'll see at least one of the blocks bouncing up and down a bit. Similarly when moving the head front to back, the front and/or rear blocks are probably also bouncing up and down. This could be caused by the pulleys being eccentric, or, more likely, bent axis rods. As the axis spins to slide the head on the opposite axis, a bent rod lifts the sliding block (and hence head) up and down - moving it closer and further away from the bed, causing the pattern you are seeing. I don't know how this affecting the top layer issue, but it probably isn't helping.

You need to make sure that the model is completely watertight and properly formed. But failing that, why not look at using a different slicer? Kisslicer, Slic3r, Simplify3D etc.

That's really odd. The tension is probably set too high in the video - it needs to be around the second line from the top on the scale. Also make sure that the filament comes easily off the spool and isn't caught up at all. What brand of material is this? Is it definitely PLA? Please double check what version of the firmware is on the printer - ideally you want 14.07 or later - see Maintenance -> Advanced -> Version. Your bed height adjustment looks like it might be a little off - the first layer isn't sticking very well, even when there does seem to be enough plastic. I think the nozzle is a touch high; you might try turning all three leveling screws about 1/8 to 1/4 turn to the left (as seen from the front of printer). But I don't think that's related to your issue.

Can you still dim the LEDs on the new printer? Looks a bit like what happens when the dimmer goes bad.

Actually, I'm not convinced that your nozzle hole is drilled off center - at least not from your photo. I just looked at a 100 nozzles, and most of them basically look like that. But it's an optical illusion caused by the threading on the column. The start of the thread cuts away part of the top surface, making it look like the hole is closer to one side of the top in most lights, because the threading reflects light differently from the top surface. If you actually screw the isolator nut down around it, and look at the alignment of the hole compared to the ring of the nut, it's actually ok.